4D Image file formats

 See https://nifti.nimh.nih.gov for a complete specification of this file format.

This format is “generally” supported on input and output.  However, there are some limitations.

  • Fastmap support NIFTI files with a self-contained header (.nii) and compressed versions of these files (.nii.gz)
  • On input, fastmap supports NIFTI_TYPE_… UINT8, INT16, INT32, FLOAT32, COMPLEX64, FLOAT64, UINT16.
  • fastmap not support image matrix dimensions above 5. Most time-series data will have dimension 4. The only currently supported 5-dimensional data is 3-vectors per voxel (e.g., vector fractional anisotropy for diffusion tensor imaging).
  • On output, fastmap will choose either short integers (NIFTI_TYPE_INT16)or floating points (NIFTI_TYPE_FLOAT32)depending upon data (e.g, short integers will be selected if all data fits into short integers). An output type of NIFTI (.nii) or gzipped NIFTI (.nii.gz) will be used depending upon GUI settings.
  • On input, fastmap also support jip input files, in which a leaderless binary file contains data specified by the extension, and an associated ASCII text file provides relevant information about the data.


Joseph B. Mandeville, Athinoula A. Martinos Center for Biomedical Imaging at MGH/MIT/Harvard